BRIDGMAN. — THERMODYNAMIC PROPERTIES OF LIQUIDS. 9 



SO viscous as not to transmit small changes of pressure hydrostatic- 

 ally immediately after changes in the position of the piston. The 

 enormously increased viscosity both of the kerosene and of the 

 liquid under investigation within the bulb afforded the only satis- 

 factory explanation of many of the capricious misfortunes of the 

 preliminary work. For instance, in the early work the bulb was 

 closed at the upper end by a cap put on with soft solder. Several 

 times after the application of pressure this soft solder was found eaten 

 away by the mercury. The onl}' apparent Avay in which this could 

 happen would be by the liquid (ethyl alcohol in the preliminary 

 investigations) becoming so viscous as to crack as the volume de- 

 creases with rising pressure, thus allowing the mercury to rise through 

 the cracks from below. Another very troublesome effect in the 

 preliminary work was the frequent short circuiting of the manganin 

 coil by small drops of mercury. These were probably forced out 

 of the cup by the viscous motion of the kerosene during changing 

 pressure. The difficulty was avoided by making the lower part of 

 the mercury cup in the form of a protecting cap for the manganin 

 coil, and by enlarging the channels of communication between the 

 cup and the bulb from the kerosene to the mercury. It was found 

 also that it was necessary to give the bulb fairly thick walls. Other- 

 wise, when pressure is relieved, the bulb expands under the viscous 

 motion of the fluid inside and tightly fills the cylinder. Furthermore, 

 the bulb must not fit the cylinder too closely, for otherwise the pres- 

 sure is not transmitted rapidly enough to the interior of the bulb, 

 which may thereby become collapsed at the upper end. Still again, 

 the upper end of the bulb must be provided with radial grooves to give 

 access from the lower to the upper part of the cylinder, or else 

 when pressure is released, the entire bulb rises with the viscous 

 kerosene against the ledge on the cylinder, acting effectively as a 

 valve which permits only comparatively slow release of pressure, 

 with the result that e\'entually, when' pressure is released, the top 

 is blown off the bulb and forced into the smaller bore of the cylinder 

 above. 



It must be a matter of experiment to find the dimensions for any 

 particular piece of apparatus which will avoid all these difficulties. 

 Complete success was attained, however; the last twenty out of 

 a total of twenty-four runs being completed without accident of any 

 kind. 



Although the method outlined above is exceedingly simple, yet there 

 are evidentlv a number of corrections for the distortion of the steel 



